5 Keys To Understanding Why Alternative Credit Scoring Changes The Game

Alternative credit scoring is one of the latest advances in financial technologies. Not only does it allow the creation of more successful credit profiles, but it also helps banks to grow their customer base by leaps and bounds.

But what exactly is this novel technique about? And how did one of the most influential banks in Indonesia start using it to grow 107% in the amount of loans granted?

Read on to find out.

More Realistic Profiles Faster Thanks To Machine Learning

How to determine if someone is a good candidate for a loan? In the traditional way, a credit profile is built using data such as remuneration and income level. The problem? Not only are the profiles created with this data just a slice of reality, but the process to create them is slow and bureaucratic. In the end you end up working hard to collect little.

Alternative scoring models, on the other hand, use machine learning to process every aspect of users’ lives to the smallest detail. Not only does it analyze traditional data, but it also collects records, examines payment capacity, analyzes online payment activity, and even takes into account social networks!

Just take a look at all the variables that a platform like Mercado Libre takes into account for its credit program:

All this being analyzed in real time through complex algorithms to determine how good a candidate for a loan is a user. The result is more realistic and much more complete profiles and more quality clients.

Allows Capture More Clients

Traditional scoring methods sometimes stunt the growth of banks. When all you have to create a profile is a small handful of numbers, it is almost impossible to know for sure if someone is a good candidate for a loan. This leads banks to reject tons of applicants who would have been ideal clients.

In the case of Mercado Libre mentioned in the previous point, these complex algorithms allowed the platform to approve 60% of credit applications with complete confidence, of which 67% are for micro-enterprises. The most interesting? 80% of these people would not have been considered “good profiles” by traditional means.

But of course, this does not only happen with fintech companies. Despite its desire to attract new clients, one of the top 10 banks in Indonesia could not avoid rejecting 85% of the requested loans. The solution? After adopting an alternative scoring method to achieve realistic profiles of its prospects, the bank began approving 107% more loans – now with the assurance that they were choosing trustworthy clients.

Machine Learning Also Allows You To Avoid Fraud

If there is something that characterizes this moment in history, it is connectivity. We are all connected. All of our information is floating on a server. Not using it is wasting the enormous potential of Big Data.

When it comes to detecting fraud, the same machine learning systems that potential alternative scoring also detect when someone may not have the best of intentions.

Components such as the SDK allow you to detect behavioral anomalies and alert you to potential fraud even before it happens. Does a user carry out suspicious activities such as trying to erase their tracks or access towering credits from unsuspected places? The system knows it.

Thanks to these anti-fraud systems, platforms such as Credolab have reduced the delinquency rate of their loans by 15% and the fraud rate by 22%.

Big Data is a powerful tool.

Fewer Prospects, More Clients

Until now, a financial institution could take up to 30 days with the bureau to obtain updated information from a client. On the other hand, with Open Banking any modernized bank can access this information easily. Thanks to this, little by little it is possible to eliminate the figure of the prospect — risky when granting loans — in pursuit of clients and new clients.

Thanks to Open Banking, an identity can access a new customer’s information simply by asking their long-time bank for all the information collected. It is a more agile process, faster, and with less chance of losing a customer.

How Does BATE Gap Influence Alternative Scoring Capability?

The BATE Gap is a concept that we developed at N5 to measure the gap between the level of banking in a country and the penetration of mobile telephony. This powerful tool allows us to determine which locations are ideal for implementing mobile banking processes.

Take a look at the following chart:

As you can see, in almost all the countries measured, the penetration of mobile telephony exceeds the levels of banking. In some it is up to double!

What does this mean? That the greater the gap measured in the BATE Gap, the better the opportunity to implement alternative scoring methods.

This means that there is currently a huge unexplored market in developing countries – as indicated by the BATE Gap – and a huge opportunity for those who enter now, before supply is saturated.

An example of this is the microcredit market that has not stopped growing since 2008, and which is expected to reach 650 million dollars in 2025.

Editorial: Marcelo Frette